VarScan

VarScan detects sequence variants (SNPs and indels) from short-read sequencing data and, in VarScan 2, identifies somatic mutations and copy number alterations in tumor–normal exome pairs.

Key Features:

• Sequencing platform support: Compatible with Roche/454 and Illumina/Solexa short-read sequencing data.
• Aligner compatibility: Works with multiple short-read aligners rather than a single aligner.
• Germline variant detection: Calls single nucleotide polymorphisms (SNPs) and insertions/deletions (indels) from sequencing data.
• Sensitivity and specificity: Demonstrates high sensitivity and specificity for variant detection in individual and pooled deep sequencing datasets.
• VarScan 2 somatic analysis: Detects somatic mutations and classifies variants in tumor–normal exome pairs.
• Paired-sample analysis: Reads tumor and normal samples simultaneously to support somatic calling and classification.
• Variant classification: Uses heuristic and statistical algorithms to classify variants as germline, somatic, or loss-of-heterozygosity (LOH) events.
• Copy number analysis: Uses normalized read depth comparisons to delineate relative copy number changes and identify large-scale and focal CNAs.
• Recurrent CNA detection: Identifies recurrent focal CNAs using the correlation matrix diagonal segmentation (CMDS) algorithm.
• Cancer genomics validation: Validated on exome data from 151 high-grade ovarian tumors in TCGA, reporting validation of ~7,790 somatic coding mutations with high sensitivity and precision.
• Biological insights: Enabled detection of recurrent oncogene amplifications and tumor suppressor deletions in cancer datasets.

Scientific Applications:

• Germline variant discovery: Detection of SNPs and indels for studies of DNA sequence variation in human disease.
• Somatic mutation discovery: Identification and classification of somatic mutations from tumor–normal exome sequencing.
• Copy number alteration analysis: Detection of large-scale and focal CNAs in tumor exomes using normalized read depth.
• Deep and pooled sequencing analysis: Variant calling in individual and pooled deep sequencing datasets from Illumina/Solexa.
• Cancer genomic profiling: Identification of recurrent focal CNAs and recurrently altered genes in cancer cohorts such as TCGA ovarian tumors.

Methodology:

VarScan applies heuristic and statistical algorithms, performs simultaneous tumor–normal read comparisons, uses normalized read depth comparisons for copy number estimation, and employs correlation matrix diagonal segmentation (CMDS) for recurrent focal CNA detection.